J.M. Howarth¹, S.P. Miller² and J.W. Wilton^2
1Animal Genetics and Breeding Unit, University of New England, Armidale NSW 2351, Australia
²Centre for the Genetic Improvement of Livestock, Animal & Poultry Science,
University of Guelph

Summary

The simultaneous selection for two breeding objectives of Ontario beef production was evaluated using two breeding strategies. In strategy A, two lines were used to select directly for each breeding objective while in strategy B a single line was selected for the average of the objectives. After 15 years of selection, strategy B was superior to strategy A for all positive correlations between breeding objectives and at the expected correlation between breeding objectives of 0.5 this advantage was maintained across the entire time horizon. Strategy B simplifies the application of multiple objective selection for the individual breeder.

Introduction

Two different economic breeding objectives have recently been defined for Ontario beef breeders by the Centre for Genetic Improvement of Livestock (CGIL). One has targeted the production of large carcass weight cattle while the other has been aimed at producing a maternal animal with higher carcass quality.

With the definition of more than one breeding objective, the problem exists of how to best improve both breeding objectives. To optimise response in the two breeding objectives the breeding herd could be divided into two lines and selected directly for each breeding objective across both lines (strategy A) or selected for the average of the objectives in a single line (strategy B). The aim of this paper was to determine which strategy would maximise the total average merit of the two breeding objectives.

Materials and Methods

Demand for seedstock of both breeding objectives was assumed to be equal so that in strategy A the lines were of equal size and in strategy B equal weight was given to each breeding objective. For a breeding herd of 100 females, selection response in each breeding objective was predicted assuming that 10 animals were selected per year from the single herd of strategy B and from each line of strategy A. Selection was applied for 25 years assuming that the variances of each objective were equal and the period between generations was 5 years. With every 5 years of selection, the animals of the breeding herd were allocated into two groups according to the breeding objective for which they had the highest value. Each strategy was measured on the total average merit of the two allocated groups.

Results and Discussion

Figure 1 shows the total average value of each strategy for the complete range of correlations between breeding objectives after 15 years of selection. In this case, strategy A is superior to strategy B when the correlation is less than -0.1 but for all correlations greater than this strategy B generates the greatest total average merit, even though the difference is small.

Preliminary analysis of the two Ontario breeding objectives has shown them to have a correlation (r_1,2) of approximately 0.5 which means that selection for one objective will result in the improvement of the other but at approximately half the rate. At this correlation the superiority of strategy B over A is at its greatest and from Table 1, it can be seen that proportionally, this advantage is relatively constant over the time horizon considered here. However, as strategy B selects on the average of the objective rather than the true objectives as in strategy A it would be expected that with further selection, strategy B would lose to this advantage.

The assumptions on which this work was based allowed general conclusions to be made on how best to select simultaneously for two breeding objectives. In reality, demand for animals selected for different breeding objectives may differ resulting in unequal rates of response in each objective. Furthermore, the current mean values for each objective are likely to be different for each breeding herd whereby the proper weighting on each objective would need to be specified for each breeder in order to maximise total average merit.

The results presented here show that selection for the average of two breeding objectives will produce greater total average merit than would a breeding strategy selecting for each objective in two lines. The additional advantages associated with strategy B are that animals are only selected for a single objective rather than two in strategy A and that the size of the breeding herd will be limited when strategy A can be applied.

Significance to the Industry

For the anticipated correlation between the breeding objectives of beef production in Ontario, selecting the average of the breeding objectives will generate total average merit that is greater than would be achieved by selecting in two different lines. The advantage of the average strategy B is not only maintained over the long term but simplifies the application of multiple objective selection for the individual breeder.

Acknowledgements

This work was made possible through the support of Beef Improvement Ontario, the Natural Sciences and Engineering Research Council of Canada, the Australian Murray Grey Beef Cattle Society and the University of New England through the Keith and Dorothy Mackay travel scholarship.

Figure 1. Comparison of total average merit across correlations between strategies A and B after 15 years of selection.